Security audit

Logo设计工具专业版

Security checks across malware telemetry and agentic risk

Overview

This logo-design skill is not clearly malicious, but it gives agents broad command, file-writing, external-service, and API-key guidance without tight boundaries.

Review before installing. Use it only in a project directory you are comfortable letting an agent modify, avoid giving it broad API credentials, require confirmation before any command execution or external AI call, and do not send confidential brand assets unless the provider and retention terms are clear.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Context-Inappropriate Capability

Medium

Confidence: 84% confidence
Finding: The skill mixes logo-design functionality with broad API credential guidance, API key pools, multi-account rotation, and external service usage without tightly scoping what services are required or what data is sent. That increases the chance an agent will request, store, or transmit secrets and user content to unnecessary third parties, expanding the attack surface beyond the stated design use case.

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: Advertising generic command-execution capability in a skill that already has exec/write tools encourages agents to run arbitrary system commands beyond the narrowly defined logo workflow. In practice, overly broad execution framing can be abused for unintended filesystem changes, environment inspection, or lateral actions unrelated to design tasks.

Intent-Code Divergence

Medium

Confidence: 81% confidence
Finding: The documentation contradicts itself by first saying no extra API key is needed and then stating AI generation tools require API keys and support key pools. Conflicting credential requirements can cause unsafe operator behavior, such as entering secrets unnecessarily, misconfiguring access, or trusting unclear authentication flows.

Vague Triggers

Medium

Confidence: 78% confidence
Finding: The invocation guidance uses broad everyday wording like general design, UI, poster, and brand-visual use, which encourages over-triggering the skill outside a tightly bounded logo-design context. Because the skill also supports exec/write and external integrations, broad activation language increases the chance it will be selected for tasks involving unnecessary file changes or data sharing.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: The skill describes command execution and file-writing behavior without a prominent upfront warning that it can create or modify files and run local scripts. Users may invoke it expecting advisory design help, while the agent instead performs system-affecting actions that could overwrite data or execute unsafe local code paths.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill references API keys, external AI tools, and load-balanced multi-account usage without warning that prompts, images, and brand assets may be transmitted to third-party services. In a commercial branding context, those materials may be sensitive or proprietary, so silent external transmission creates real confidentiality and compliance risks.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.